Upload
others
View
2
Download
0
Embed Size (px)
Citation preview
Filed: 2013-12-19
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 1 of 49
DISTRIBUTION BUSINESS LOAD FORECAST AND 1
METHODOLOGY 2
3
4
1.0 INTRODUCTION 5
6
This exhibit discusses Hydro One Distribution’s system load forecast and methodology. It 7
provides information on a distribution total basis that assists Hydro One Distribution in 8
forecasting the work programs that need to be undertaken by Hydro One Distribution to meet 9
customers’ electricity demands, and to accommodate new customer connections. 10
11
Hydro One Distribution uses a number of methods, such as econometric models, end-use 12
models, and customer forecast surveys to produce the forecasts required for its distribution 13
business. Similar methods are used by major utilities throughout North America. 14
15
All forecasts presented in this section are weather-normal and the numbers are at the 16
wholesale level unless otherwise specified. Abnormal weather effects are removed from the 17
base year for load forecasting purposes so that the forecast assumes typical weather 18
conditions based on the average of the last 31 years. The weather correction methodology 19
used by Hydro One Distribution is a proven technique that has performed well in past years. 20
The same methodology was reviewed and approved by the Board in the Distribution Cost 21
Allocation Review (EB-2005-0317) and for Hydro One’s previous Distribution Rate cases 22
(RP-2005-0020/EB-2005-0378, EB-2007-0681, and EB-2009-0096). 23
24
All forecasts produced are internally consistent meaning all customer groups add up to the 25
total customer base served by Hydro One Distribution. 26
27
Filed: 2013-12-19
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 2 of 49
Hydro One Distribution’s load forecast staff has significant experience in preparing 1
provincial and local electricity demand forecasts and load profiles. The load forecast 2
methodology described in this exhibit is the same as Hydro One’s previous Distribution Rate 3
cases (RP-2005-0020/EB-2005-0378, EB-2007-0681, and EB-2009-0096). The performance 4
of Hydro One Distribution’s system load forecast, since Hydro One Distribution’s separation 5
from the former Ontario Hydro, has been accurate as shown in Table 1. 6
7
Between 1997-2001, the average variance of customers’ energy purchase forecast compared 8
to the weather corrected actual energy consumed is within one standard deviation of the 9
forecast, despite large variances resulting from unusual events such as the Ice Storm in 1998 10
and September 11 in 2001. One standard deviation, an accepted standard in the utility 11
industry, means there is one in three chances that the actual will be outside the plus or minus 12
range (alternatively, there is two in three chances that the actual will fall within the plus or 13
minus range). The performance of the forecast in subsequent years, namely 2002 to 2012, 14
shows that the forecast is tracking very well and certainly well within one standard deviation 15
band for the corresponding energy purchases. 16
17
Table 2 compares the accuracy of the load forecast approved in the last Distribution Rate case 18
(EB-2009-0096) with the weather corrected actuals. Detailed forecast accuracy comparisons 19
of the previous three Hydro One Distribution rate applications (EB-2005-378, EB-2007-0681 20
and EB-2009-0096) with the weather corrected actuals are presented in Appendix E, Table 21
E.1.22
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 3 of 49
Table 1 1
Comparison of Hydro One Distribution Forecast with Actual 2
(Variance of forecast expressed as percent of actual on weather corrected basis) 3
____________________________________________________________________ 4
Forecast made Variance for Variance Variance 5
for Plan Year Plan Year for 2nd
Year for 3rd
Year 6
______________________________________________________________________ 7
1997 0.12 -2.03 1.91 8
1998 -2.03 -3.39 -2.02 9
1999 -0.85 0.73 -0.15 10
2000 0.46 -0.03 0.76 11
2001 -1.80 -1.56 -2.44 12
2002 1.98 2.39 2.12 13
2003 -0.82 -1.37 -0.74 14
2004 0.14 0.62 0.76 15
2005 0.25 0.12 0.46 16
2006 -0.06 -0.12 0.99 17
2007 -0.09 0.93 1.59 18
2008 -0.57 0.54 0.70 19
2009 -0.14 -0.25 -0.78 20
2010 1.24 0.28 -0.73 21
2011 0.22 0.34 N/A 22
2012 0.54 -0.51 N/A 23
2013 -0.39 N/A N/A 24
____________________________________________________________________ 25
Mean (1997-2001) -0.82 -1.26 -0.96 26
One standard deviation (+/-) 1.13 2.57 3.00 27
Mean (2002-2013) 0.19 0.27 0.41 28
One standard deviation (+/-) 1.07 2.42 2.79 29
______________________________________________________________________ 30
Note: The forecast performance pertains to Hydro One Retail purchases, which account for about 96 percent of 31
the revenue requirements in the Distribution Rate case. 32
33
34
Filed: 2013-12-19
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 4 of 49
Table 2 1
2
3
4
5
Section 2 provides a more detailed discussion in respect of the various economic 6
considerations that Hydro One Distribution staff take into consideration when applying the 7
methodology for deriving the load forecasts. 8
9
Hydro One Distribution’s forecasting methodology uses a combination of elements that 10
include consensus input, updates to changes in economic forecasts, energy prices, population 11
and household trends, industrial development and production, residential and commercial 12
building activities, and efficiency improvement standards. Economic inputs were based on 13
analyses prepared by major economic establishments in the country such as IHS Global 14
Insight, Conference Board of Canada, Centre for Spatial Economics, University of Toronto, 15
Canada Mortgage and Housing Corporation, and Altus Group. 16
17
Efficiency standard assumptions used in the end-use models are based on discussion with 18
Ontario Ministry of Energy staff. Specific customer development is based on forecast survey 19
results from major customers. Inputs from these entities form the economic database 20
(referred to henceforth as economic forecast) that is used to establish Hydro One Distribution 21
load forecast. Section 3 below provides a detailed description of the methodology used by 22
Comparison of 2009 Forecast with Actual
(GWh)
Weather
Retail Corrected
Year Forecast Actual Variance (%)
2009 22,629 22,660 -0.14
2010 22,007 22,062 -0.25
2011 21,851 22,023 -0.78
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 5 of 49
Hydro One to develop its load forecasts. Detailed modeling equations and definitions are 1
presented in the Appendices. Details of the consensus forecast for Gross Domestic Product 2
(GDP) and housing starts are provided in Appendix E, Table E.1. 3
4
Using Hydro One Distribution’s approved forecasting methodology, the forecast for the test 5
years (2015 – 2019) is presented below: 6
Year GWh Delivery
Forecast
Distribution Customer
Count
2015 37,620 1,288,000
2016 37,824 1,300,000
2017 38,108 1,312,000
2018 38,111 1,325,000
2019 37,961 1,337,000
7
The 2015 figures represents an increase of 0.3 percent over the 2013 load forecast and an 8
increase of 1.6 percent over the 2013 customer count. The small increase in load is mainly 9
due to the impact of Conservation and Demand Management (CDM) and the current 10
economic conditions. Section 4 provides a more detailed discussion in respect of the 11
comparison of the test years (2015-2019) forecasts in relation to the historic (2012 and 2013) 12
and bridge (2014) year. 13
14
This Exhibit addresses two directives from the Board’s April 9, 2010 Decision on Hydro 15
One’s Distribution rate application for the 2010 and 2011 years (EB-2009-0096) requiring 16
Hydro One to: 17
18
1) track the difference between the CDM forecast assumed in the load forecast and CDM 19
impacts actually achieved in 2010 and 2011; and20
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 6 of 49
2) to provide a detailed analysis for estimating the CDM impacts and to develop a methodology to 1
incorporate these impacts into the load forecast. 2
3
In response to the first directive to track the CDM impacts, a detailed report tracking the 4
CDM results for the 2005-2013 period was prepared and is provided in Exhibit A, Tab 16, 5
Schedule 3. With respect to the second directive, Hydro One consulted with stakeholders and 6
worked with the Ontario Power Authority (OPA) to use the latest CDM assumptions in 7
preparing the load forecast in this rate application. A detailed report was prepared and is 8
provided in Exhibit A, Tab 16, Schedule 4. Summary results of these 2 reports are discussed 9
in Section 2.6. 10
11
2.0 DISCUSSION OF THE ECONOMIC CONSIDERATIONS THAT 12
INFLUENCE HYDRO ONE DISTRIBUTION’S LOAD FORECASTS 13
14
This section discusses some of the key economic considerations that must be taken into 15
account in the process of developing load forecasts and in the application of forecasting 16
methodologies. The elements of the forecasting process used by Hydro One Distribution are 17
for the most part based on the knowledge of how the major economic drivers that affect the 18
usage of electricity demand are likely to occur over the forecast period (2015 to 2019). 19
Consequently for the purpose of this application the focus is on the short term and medium 20
term and the load forecast will reflect those impacts that are likely to have a major effect in 21
this respect. The major economic drivers used in the analysis are summarized in Figure 1. 22
23
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 7 of 49
1
2
Key information used in the analysis includes the Ontario GDP, provincial demographic, 3
industrial production and commercial output forecasts and regional analysis included in the 4
economic forecast. Also taken into consideration are Hydro One Distribution CDM plans, 5
which have a direct impact on distribution system energy demand. 6
7
The load forecast in support of this application was updated in April 2014 using economic 8
information and forecasts that were available in March 2014. The update included the latest 9
economic forecast, 2013 actual purchases, and CDM detailed information released by the 10
OPA consistent with the 2013 Long-Term Energy Plan (LTEP). The timing of the load 11
forecast is driven by the needs of the business planning process which are geared to match 12
the timeline for this submission. 13
14
2.1 Provincial GDP Forecast 15
16
The provincial GDP forecast is a key driver for the load forecast. The Ontario GDP grew 3.4 17
percent in 2010 and then slowed to 2.2 percent in 2011 and 1.3 percent in 2012. The Ontario 18
Figure 1Hydro One Distribution Load Forecast Methodology
Key Drivers Hydro One Distribution Load Forecast
- Provinical GDP forecast Econometric Approach End-use Approach - Population and household forecast - Monthly model - Forecast by sector and by - Housing forecast - Annual model end-use - Industrial production forecast - Commercial output forecast
Key Drivers Sub-Transmission Customer Load Forecast
- Provinical GDP forecast Forecast by customer - Population and household forecast - Econometric analysis - Housing forecast - Analysis by customer - Industrial production forecast - Customer survey results
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 8 of 49
economy is expected to grow slowly in the next 2 years. Based on the consensus forecast, the 1
Ontario economy is projected to grow by about 1.2 percent in 2013 and about 2.2 percent in 2
2014. Over the 2015-2019 period, the provincial GDP is forecast to grow on average 2.6 3
percent per year. Details of annual forecast for GDP, population and housing starts are 4
provided in Appendix E, Table E.3. 5
6
2.2 Provincial Population Forecast 7
8
Ontario population grew on average by about 1.1 percent per year between 2010 and 2012. 9
Population growth for the province is forecast to outperform the nation in the forecast period. 10
Ontario population is expected to grow by about 0.9 percent per year over the 2013-2014 11
period and by about 1.0 percent per year over the 2015-2019 period. Steady population 12
growth contributes positively to the load forecast. 13
14
2.3 Provincial Housing Forecast 15
16
Helped by historically low interest rates, housing starts remained strong in the last few years, 17
growing on average about 68,000 units per year between 2010 and 2012. Due to the 18
continued slow economic recovery, housing starts are expected to slow to about 61,000 units 19
in 2013 and 59,000 units in 2014, and then recover over the 2015-2019 period, averaging 20
about 69,000 units per year. 21
22
2.4 Commercial Output Forecast 23
24
Commercial activities follow closely with the general economic conditions. After growing 25
2.8 percent in 2010, the commercial output slowed to 1.8 percent in 2011, 1.4 percent in 26
2012, and 1.6 percent in 2013. The growth is expected to continue in 2014 at the rate of 2.6 27
percent. Commercial output is projected to have moderate growth over the 2015-2019 28
period, averaging about 2.8 percent growth per year. Commercial output is important to the 29
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 9 of 49
load forecast because commercial load comprises about 25 percent of Hydro One 1
Distribution’s load. 2
3
2.5 Industrial Production Forecast 4
5
After 4 years of decline, the industrial production grew 7.0 percent in 2010, followed by 4.4 6
percent in 2011 and 1.3 percent in 2012, consistent with the slow recovery pattern of the 7
world economy. Due to the high Canadian dollar and slow economic growth, the industrial 8
output declined by 2.4 in 2013. The forecast calls for slow growth of 0.6 percent in 2014 and 9
moderate growth over the 2015-2019 period, averaging about 2.5 percent per year. The 10
industrial production forecast is important to the load forecast because industrial activity 11
comprises about 10 percent of total load and also because it is prone to the impact of 12
economic cycles. 13
14
2.6 Conservation and Demand Management 15
16
The Board in its EB-2009-0096 Decision directed Hydro One Distribution to track the 17
differences between its CDM forecast and those which can be reasonably demonstrated to 18
have been affected using the best verification methods available at the time of filing. In 19
response to this directive, Hydro One completed a detailed study entitled “Tracking 2005-20
2013 Conservation and Demand Management Results” and is provided in Exhibit A, Tab 16, 21
Schedule 3. Hydro One has undertaken extensive research and analysis to assess the CDM 22
impacts for use in its load forecast. The results of the tracking study show that the CDM 23
impact assumed in Hydro One Distribution’s load forecast for 2010 to 2011 approved by the 24
Board in EB-2009-0096 was appropriate. As demonstrated in Table 2 in this Exhibit, Hydro 25
One’s load forecast for retail customers also tracks well with the actuals on a weather-26
corrected basis for 2010 and 2011. 27
28
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 10 of 49
In the tracking study, the CDM impacts on Hydro One Distribution load are grouped in the 1
following categories: 2
3
CDM Programs initiated by both Hydro One and the Ontario Power Authority; 4
CDM programs funded by other organizations, such as federal, provincial, and/or 5
municipal governments, natural gas companies, and other non-government organizations; 6
CDM impacts from Building Codes and Appliance Efficiency Standards (approved after 7
2004); and 8
CDM Impacts from Increased Conservation Effect not captured by specific CDM 9
programs. This category includes actions which consumers take to improve efficiency 10
and conservation above and beyond historic levels of “naturally occurring conservation” 11
due to increased awareness of, and concern about, environmental or energy issues. 12
13
The top-down and bottom-up analyses presented in the tracking study show that Hydro One 14
Retail customers have responded to the conservation challenge, have participated in CDM 15
programs offered by the OPA, Hydro One and other government agencies, and have taken 16
various conservation actions on their own to save electricity. 17
18
In its EB-2009-0096 Decision, the Board restated its directive to Hydro One in EB-2007-19
0681 to come forward in its next cost of service application with a detailed proposal to 20
incorporate the impacts of CDM in its load forecast, both attributable to its own actions and 21
those attributable to other factors. In its EB-2010-0002 Decision for Hydro One’s 22
Transmission rate application, the Board also directed Hydro One to consult with the 23
stakeholders and work with the OPA to devise an effective and accurate means of measuring 24
the expected impacts of CDM programs delivered by the OPA. Hydro One worked with the 25
stakeholders and listened to and addressed their concerns. Hydro One also worked with the 26
OPA and obtained their current CDM forecast consistent with the 2013 LTEP for use in this 27
rate application. Details of the information provided by the OPA and the methodology used 28
Filed: 2013-12-19
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 11 of 49
by Hydro One to derive the CDM impacts by rate class are documented in a detailed report 1
entitled “Incorporating Conservation and Demand Management in the Distribution Load 2
Forecast” and is provided in Exhibit A, Tab 16, Schedule 4. The CDM forecasting study 3
focuses on two objectives: 4
(i) propose a methodology to incorporate CDM impacts into the load forecast; and 5
(ii) derive CDM impacts for use in Hydro One’s distribution load forecast. It should be 6
noted that the content of this report is similar to the CDM forecasting report filed in 7
Hydro One’s last Transmission Rates Application (EB-2012-0031, Exhibit A, Tab 15, 8
Schedule 2, Attachment 1) which was approved by the Ontario Energy Board. For 9
convenience, the information is re-submitted in this report and updated where 10
appropriate. 11
12
Table 3 summarizes the CDM impact assumed in Hydro One’s distribution system load 13
forecast. Details of CDM forecast by rate class are provided in Appendix E, Table E.9. 14
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 12 of 49
Table 3 1
CDM Impact on Hydro One Distribution Load 2
(GWh) 3
4
Retail ST Customers
Year Customers Direct LDC Total
2012 1,237 412 704 2,353
2013 1,284 421 789 2,494
2014 1,336 426 893 2,655
2015 1,374 427 968 2,769
2016 1,417 429 1,000 2,845
2017 1,416 420 1,009 2,846
2018 1,646 454 1,154 3,253
2019 1,949 505 1,342 3,796
Note. All figures are weather-normal.
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 13 of 49
2.7 Customer Forecast 1
2
Through its distribution system, Hydro One is expected to serve about 1.268 million 3
customers in 2013 and 1.279 million customers in 2014. The customer base is forecast to 4
reach 1.288, 1.300, 1.312, 1.325, and 1.337 million respectively over the 2015-2019 period. 5
Detailed customer information is retained in the customer information system for billing and 6
account management. Customer data is extracted from the system regularly for tracking, 7
analysis and reporting. The customer forecast was developed on an as-required basis to 8
support the annual business planning process, system development plans and rate 9
submissions to the Board. Active customer accounts and service points are used as the basis 10
to prepare the customer forecast by rate class. The customer forecast takes into consideration 11
the new customers requiring distribution services, existing customers moving out, provincial 12
housing demand, population and household forecasts, vacancy rates and specific growth 13
patterns of various customer groups. 14
15
Customer growth in Hydro One Distribution averaged about 9,000 per year over the 2009-16
2012 period, which is lower than normal growth due to the recent 2008-2009 economic 17
downturn and the slow recovery afterwards. During the 5 years prior to 2009, customer 18
growth averaged 12,000 per year. Customer growth for the forecast period is expected to 19
gradually return to the historical normal growth pattern. Net customer additions are forecast 20
to be about 11,000 per year over the 2013-2014 period and about 12,000 per year over the 21
2015-2019 period. Details of customer forecast by rate class over the forecast period are 22
provided in Appendix E, Table E.4. 23
Filed: 2013-12-19
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 14 of 49
1
3.0 LOAD FORECASTING METHODOLOGY 2
3
Hydro One Distribution’s load forecast is developed using both econometric and end-use 4
approaches. The load impacts of CDM are added back to the historical values during the 5
modeling process (see Figure 2 below). 6
Figure 2 7
Incorporation of CDM in the Load Forecast 8
9
The forecast base-year is corrected for abnormal weather conditions and the forecast growth 10
rates are applied to the normalized base-year value. Thus the forecast is weather-normal in 11
the sense that it predicts the future load under normal weather conditions. 12
13
3.1 Weather Correction Analysis 14
15
Weather correction is a statistical process designed to remove the abnormal or extreme weather 16
effects from the load data to yield average conditions that reflect the more normal or expected 17
weather conditions experienced over 31 years used in the forecast. It is essential that abnormal 18
and extreme weather related impacts are removed before establishing the base-case load data, on19
2004 2012
A
D
C
B
E
2019
Historical CDM
Projected CDM
A: 2004 actual load
B: Estimated Load without
CDM impacts in 2012
C: 2012 actual load
Filed: 2013-12-19
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 15 of 49
which basis the load forecast will be developed. The volatility of abnormal or extreme weather 1
conditions can adversely impact the ability to provide a consistent and meaningful forecast for 2
load growth. Hourly load data and hourly weather data of various weather stations across the 3
province are used in the analysis. 4
5
Hydro One Distribution’s weather correction methodology was developed jointly by 6
forecasting and meteorology staff of the former Ontario Hydro. This weather correction method 7
has been used to forecast the total system load since 1988 and for forecasting local electric 8
utility load since 1994. The weather correction methodology used by Hydro One Distribution is 9
a proven technique that has performed well in the past years. The same methodology was 10
reviewed and approved by the Board in the Distribution Cost Allocation Review (EB-2005-11
0378) and in previous Hydro One Distribution Rate applications (RP-2005-0020/EB-2005-12
0378, EB-2007-0681, and EB-2009-0096). 13
14
As shown in Table 4, using a fewer number of years for historic weather normalization has 15
only a small impact on the total weather corrected energy consumption. This is an expected 16
outcome since weather normalization has a more significant impact on peak than it does on 17
energy due to the fact that energy consumption is less sensitive to short-term weather 18
conditions. In Hydro One Distribution’s Rate case, weather normal energy (and not peak) is 19
a key measure for the load forecast. 20
21
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 16 of 49
Table 4 1
Comparison of Different Time Periods Used for Weather Normalization 2
(GWh) 3
4
5
Hydro One Distribution’s weather correction methodology uses four years of daily load and 6
weather data to establish a sound statistical relationship between weather and load at the 7
applicable transformer station or delivery point used to supply customer demand. Weather 8
variables used in the analysis include temperature, wind speed, cloud cover and humidity. The 9
estimated weather effects are then aggregated up to the required time interval. Past experience 10
shows that weather correction should best be done on a daily basis, rather than weekly, monthly 11
or annual basis. 12
13
Daily weather-correction is preferred because the timing of extreme temperatures combined 14
with wind speed and humidity can have a substantial impact on load that would otherwise not be 15
captured by averages over longer period of time. In particular, when abnormal weather 16
conditions continue for several days, the cumulative impact is much greater than would be the 17
case if the same weather conditions prevailed over a much longer period of time. 18
Actual Load Weather Correction Weather Corrected
Number of Years for Hydro One Required for Hydro One Load for Hydro One
Used to Calculate Retail Customers Retail Customers Retail Customers
Normal Weather in 2013 in 2013 in 2013
Last 31 Years * 20,668 -230 20,439
Last 20 Years 20,668 -238 20,430
Last 10 Years 20,668 -198 20,471
* Used by Hydro One Distribution to normalize the base year (2013) load.
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 17 of 49
The loads that are most impacted by changes in weather conditions are electric space heating 1
and cooling in residential and commercial buildings. Across Ontario, the penetration rate of 2
such loads varies widely, which means the weather sensitivity of load supplied from one 3
transformer station or delivery point may differ quite significantly from that of load supplied 4
from another transformer station or delivery point, even in the same climate zone. The climate in 5
Ontario varies considerably from the Niagara Peninsula to Thunder Bay, so it is important to use 6
data from the appropriate weather stations that are in close proximity to the transformer station 7
or the customer delivery point when correcting for weather effects. Forecast using 31-year, 20-8
year and 10-year weather-normalization is presented in Table 5 below. 9
10
Table 5 11
Load Forecast for Hydro One Retail 12
Using Different Normalization Periods 13
(GWh) 14
15
16
3.2 Hydro One Distribution Forecasting Methodology 17
18
Both econometric (top-down) and end-use (bottom-up) models are used to prepare load 19
forecast for Hydro One Distribution. Both monthly and annual econometric models are used 20
to forecast Hydro One Distribution’s total distribution system load. End-use models using 21
the results from the provincial end-use models are used to analyse the distribution system 22
load by customer rate class (i.e. various residential and general service customers). Key 23
information used in the analysis includes economic, demographic, industrial production and 24
commercial output forecast provided in the economic forecast. The purpose of using both the 25
econometric and end-use forecast models is to arrive at a balanced forecast that represents a 26
2015 2016 2017 2018 2019
31 years 20,497 20,630 20,808 20,825 20,759
20 years 20,488 20,621 20,799 20,816 20,750
10 years 20,529 20,662 20,840 20,857 20,791
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 18 of 49
consistent set when looked at from macro (econometric) and micro (end-use) perspectives. 1
The load impacts of CDM are added back to the historical data set during the modelling 2
process. 3
4
Monthly Econometric Model 5
The monthly econometric model uses a multivariate time series approach to develop the 6
monthly forecast for the Distribution system load. The model links monthly energy 7
consumption to Ontario GDP and residential building permits. Appendix A provides the 8
detailed regression equations and definitions. 9
10
Annual Econometric Model 11
The annual econometric model uses personal disposable income per household, relative energy 12
price and cooling and heating degree-days to prepare the forecast. Appendix B provides the 13
detailed regression equations and definitions. 14
15
End-Use Model 16
The end-use models cover the residential (year round and seasonal), commercial, industrial and 17
agricultural sectors. Detailed equations of the end-use models are provided in Appendix C. 18
19
The above models are used to prepare forecast for the following 13 rate classes: 20
21
Urban residential (high density) 22
R1 Residential, medium density 23
R2 Residential, low density 24
Seasonal 25
Urban general service, energy-billed 26
Urban general service, demand-billed 27
General service, energy-billed28
Filed: 2013-12-19
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 19 of 49
General service, demand-billed 1
Sub-transmission 2
Street lighting 3
Sentinel lighting 4
Unmetered scatter load 5
Distributed generation 6
7
3.3 Methodology for Sub-Transmission Customers 8
9
This section discusses the load forecasting methodology used for the Sub-Transmission (ST) 10
customers. These are the embedded customers who are directly connected to Hydro One’s 11
ST system or have a delivery point embedded in Hydro One’s distribution service territory 12
and include distribution utilities, industrial and commercial customers. Both econometric and 13
Filed: 2013-12-19
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 20 of 49
customer analysis based on survey results from the customers, when available, are used in the 1
forecast. This is supplemented by the economic data provided in the economic forecast. 2
3
In 2013, Hydro One Distribution conducted a customer load forecast survey with the 4
embedded distribution utilities and embedded industrial customers with more than 5 MW of 5
loads. In addition to questions relating to the total load of the customer, information at each 6
of the delivery points was also collected. The customer survey results are used in preparing 7
the customer forecast. 8
9
For embedded distribution utility customers, econometric analysis is used to prepare the load 10
forecast as a group. For industrial customers, several information sources are used to prepare 11
the forecast. These include: 12
13
historical load profile of the customer; 14
knowledge of the customer through industry monitoring; 15
forecast provided by customer through the survey; 16
company information through Hydro One Distribution account executives, industry and 17
company forecasts from industry associations and government agencies; and 18
production and industry forecasts provided in the economic forecast. 19
20
The econometric approach was used to forecast the load for embedded utilities and industrial 21
analysis was used to forecast the load for the embedded industrial customers. In both cases, 22
results from customer survey, when available, were taken into account in developing the 23
forecast. 24
25
26
27
Filed: 2013-12-19
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 21 of 49
3.4 Methodology for Hourly Load Profiles 1
2
This section discusses the methodology for generating the hourly load profiles by customer 3
class and for specific customer delivery points. 4
5
Hourly Load Shape by Rate Class 6
7
The Electricity Power Research Institute (“EPRI”)’s Hourly Electric Load Model (“HELM”) 8
was used to develop the hourly load shape for each rate class, taking out abnormal weather 9
effects and load patterns. Actual 2012 hourly smart meter data from the IESO and interval 10
meter data from our customer information system were used as a basis to develop the hourly 11
load shapes. For rate classes that hourly data was not available for all customers, the hourly 12
data was scaled to add up to the actual load for that rate class in 2012. Similarly, the hourly 13
forecast for each rate class adds up to annual forecast for that rate class. Consequently, the 14
forecast takes into account the share of each rate class in the total load and its dynamics over 15
time. In particular, the load profiles for the years 2015-2019 take into account shifts between 16
rate classes in accordance with the annual forecast. Appendix D provides more details for the 17
methodology used by Hydro One to weather-normalize the total utility load and for each rate 18
class. 19
Hourly Load Shape by Customer Delivery Point 20
21
Similarly, the HELM is used to normalize the hourly load for each of the customer delivery 22
points, taking out abnormal weather effects and load patterns. The customer forecast is used 23
to drive the customer delivery point forecast. Key information used in the analysis includes 24
hourly load and weather data. 25
26
The most up to date customer totalization table is used to retrieve hourly electricity demand 27
data for each of the customer delivery points connected to the Sub Transmission (ST) system. 28
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 22 of 49
The totalization table reflects the latest records from Hydro One Distribution. For each 1
customer delivery point, at least one full year of hourly data is retrieved and checked for data 2
quality. Hourly weather data is also retrieved to prepare weather sensitivity analysis. 3
Weather data analyzed includes temperature, wind speed, cloud cover and humidity. Data for 4
five weather stations across Ontario are used in the analysis. They include Toronto, Windsor, 5
Ottawa, North Bay and Thunder Bay. Each delivery point is linked to the closest weather 6
station. 7
8
In preparing the database for the load shape analysis, missing values are estimated by load on 9
a similar day and hour during the same month. For weather-sensitive load, weather 10
conditions are also taken into account in estimating the missing values. To perform the latter 11
task, an hourly regression model (relating load to weather conditions) for each delivery point 12
with missing values was developed. 13
14
EPRI’s HELM is used to prepare the hourly weather response analysis by each delivery 15
point. The model takes into account differences in load depending upon time of use (i.e. 16
weekdays, weekends and holidays) and weather conditions. Load of industrial customers is 17
assumed to be insensitive to weather and as such is forecast in relation to load on a similar 18
day and hour during the historical period. 19
20
4.0 LOAD FORECAST FOR 2015-2019 21
22
Hydro One’s distribution system delivered a total of 37,419 GWh in 2012 and 37,496 GWh 23
in 2013 on a weather-normal basis. Table 6 presents the load forecast before and after 24
deducting the impact of CDM. 25
26
Before deducting the impact of CDM, Hydro One’s distribution system is forecast to deliver 27
40,163 GWh in 2014. Load before deducting the CDM impacts over the years 2015 to 2019 28
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 23 of 49
is expected to continue to grow with the load forecast reaching 40,389, 40,670, 40,953, 1
41,365, and 41,757 GWh respectively on a weather-normal basis. The forecast reflects the 2
current slow economic recovery in 2014 and the moderate economic growth over the 2015-3
2019 period. 4
5
Hydro One Distribution served about 1,260,000 in 2012 and 1,268,000 customers in 2013, 6
and is forecast to serve about 1,279,000 customers in 2014. The customer numbers are 7
forecast to be 1,288,000, 1,300,000, 1,312,000, 1,325,000, and 1,337,000 respectively over 8
the 2015-2019 period. 9
10
After removing the impact of CDM, Hydro One Distribution’s load grows from 37,419 GWh 11
in 2012 to 37,496 GWh in 2013 and is forecast to grow to 37,508 GWh in 2014 on a weather-12
normalized basis. Over the years 2015 to 2019, the weather-normalized total distribution load 13
is forecast to be 37,620, 37,824, 38,108, 38,111, and 37,961 GWh respectively. Detailed 14
tables for actual and weather-normalized total load, energy and peak by rate class are 15
provided in Appendix E, Tables E.5 to E.9. 16
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 24 of 49
Table 6 1
Hydro One Distribution Load Forecast Before and After Deducting CDM Impact 2
(GWh) 3
4
5
Retail Embedded
Year Customers Customers Total
Load Forecast Before Deducting Impact of CDM
2012 21,670 18,102 39,772
2013 21,723 18,283 40,006
2014 21,749 18,414 40,163
2015 21,871 18,518 40,389
2016 22,046 18,623 40,670
2017 22,224 18,729 40,953
2018 22,471 18,894 41,365
2019 22,708 19,049 41,757
Load Impact of CDM
2012 1,237 1,117 2,353
2013 1,284 1,210 2,494
2014 1,336 1,319 2,655
2015 1,374 1,395 2,769
2016 1,417 1,429 2,845
2017 1,416 1,429 2,846
2018 1,646 1,608 3,253
2019 1,949 1,847 3,796
Load Forecast After Deducting Impact of CDM
2012 20,434 16,985 37,419
2013 20,439 17,073 37,512
2014 20,413 17,095 37,508
2015 20,497 17,123 37,620
2016 20,630 17,194 37,824
2017 20,808 17,300 38,108
2018 20,825 17,286 38,111
2019 20,759 17,203 37,961
Note. All figures are weather-normal.
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 25 of 49
Since the forecast is weather-normal, the actual load could be below or above the forecast 1
depending on the weather conditions and/or a different economic growth pattern. Table 7 2
presents the upper and lower bands of one standard deviation for the Hydro One Distribution 3
system load forecast. Based on historical data, there is a two in three chance that the actual 4
over the forecast years (2013-2019) will fall within the upper and lower bands. The bands 5
are derived using a Monte Carlo simulation technique relating variations in load to variations 6
in Ontario GDP and weather. 7
8
Table 7 9
One Standard Deviation Uncertainty Bands for Hydro One Distribution Load 10
(GWh) 11
12 Note: 2012 and 2013 are Actuals 13
14
15
Year Lower Bound Forecast Upper Bound
2012 37,419 37,419 37,419
2013 37,496 37,496 37,496
2014 36,622 37,508 38,408
2015 36,570 37,620 38,676
2016 36,508 37,824 39,128
2017 36,529 38,108 39,743
2018 36,221 38,111 40,073
2019 35,837 37,961 40,243
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 26 of 49
APPENDIX A 1
MONTHLY ECONOMETRIC MODEL 2
3
The monthly econometric model uses the State-Space approach in the regression equation, 4
where the left-hand side of the equation represents the energy estimates, and the right-hand side 5
contains the explanatory variables including the dummy variables that are used to capture 6
special events that could affect the energy estimates because these events would likely cause 7
variations in the load. The dummy variables are used to minimize the variability of the energy 8
estimates around the forecast. 9
10
LRTLT = f (LGDPONT, LBPONT, D98Jan) 11
where: 12
LRTLT = logarithm of Distribution load, 13
LGDPONT = logarithm of Ontario GDP in constant 1997 dollars, 14
LBPONT = logarithm of Ontario residential building permits in constant dollar, 15
D98Jan = dummy variable to account for the load impact of 1998 Ice Storm, equals 1 in 16
January 1998 and zero elsewhere, 17
18
The output parameters from the model are presented below. The State-Space (SS) estimated 19
parameters are not associated with standard error and t-ratios (statistical relevance test). 20
21
State-Space (SS) 22
Seasonal Factors parameters: 23
A[1] -0.152559 24
K[1] -0.519685 25
26
Non-Seasonal 27
Factors SS parameters: 28
A[1] 0.464085 29
K[1] -0.360868 30
GDPONT[-4] 0.0588351 31
BPONT[-8] 0.00607621 32
D98JAN -0.0184342 33
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 27 of 49
R-squared = 0.988, R-squared corrected for mean = 0.988, Durbin-Watson Statistics = 2.24. 1
The goodness of fit, or the extent to which variability in the energy estimates is captured in the 2
forecast, is measured in terms of R-squared (adjusted for mean), which in this case is close to 1. 3
This result reflects statistical significance of the explanatory variables that are used to explain 4
for the variations in load. In fact, the results show that in this case the fit is very good, and 5
therefore there is confidence that the forecast will produce outcomes that are within the expected 6
range of variability. 7
8
Using the forecast values for GDP, building permits and dummy variables, the above parameters 9
are used in the monthly regression equation described on the previous page to generate the 10
forecast for Hydro One Distribution load. 11
12
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 28 of 49
APPENDIX B 1
ANNUAL ECONOMETRIC MODEL 2
3
Annual econometric model uses personal disposable income per household, relative energy 4
price, and heating degree-days to prepare the forecast. The annual model is expressed in the 5
following regression equation: 6
7
LRTLT=C(1)+C(2)*LYPDPHH+C(3)*(LPELRES-LPGASRES) 8
+C(4)*LHDD+C(5)*LRTLT(-1)-C(4)*C(5)*LHDD(-1)+C(6) 9
*D99A+C(7)*TR+C(8)*TR2+C(9)*D08ON 10
where: 11
LRTLT = logarithm of Distribution load, 12
LYPDPHH = logarithm of Ontario personal disposable income per household in constant $, 13
LPELRES = logarithm of electricity price for Ontario residential sector, 14
LPGASRES = logarithm of natural gas price for Ontario residential sector, 15
LHDD = logarithm of heating degree days for Pearson International Airport, 16
D99A = dummy variable to account for annexation of retail customers by municipal utilities 17
equals 1 after 1999 and zero elsewhere, 18
TR = a dummy variable to account for a shift in growth pattern of Distribution load, 19
increases by 1 per year prior to 1989 and no increase afterwards, 20
TR2 = TR to power 2, 21
D08ON = a dummy variable to account for economic changes, equals zero prior to 2008 and 1 22
elsewhere. 23
C(1) – C(9) = variable coefficients. 24
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 29 of 49
The estimated coefficients and associated statistics are presented below: 1
2
Estimated
Coefficient
Standard
Error
t-ratio
C(1) 7.613530
1.270369
5.993166
C(2) 0.465765
0.087124
5.345983
C(3) -0.031541
0.011513
-2.739600
C(4) 0.001654
0.029618
0.055857
C(5) 0.147403
0.098630
1.494506
C(6) -0.041241
0.008456
-4.877124
C(7) -0.124739
0.023341
-5.344296
C(8) 0.003182
0.000540
5.892496
C(9) -0.017782
0.007137
-2.491550
R-squared = 0.995, Adjusted R-squared = 0.994, Durbin-Watson Statistic = 2.20. 3
4
Similar to the regression analysis in the case of the Monthly Econometric model above, the 5
goodness of fit, measured by (Adjusted) R-square for the Annual Econometric Model, is also 6
found to be close to 1. Therefore the assessment on an annual basis also leads to a forecast 7
outcome which provides consistent results, thus giving confidence to the econometric method. 8
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 30 of 49
The t-ratios show most of the factors used to explain the variations in load are statistically 1
significant. 2
3
Using the forecast values for personal disposable income, energy prices, and heating degree 4
days and dummy variables, the above parameters are used in the annual regression equation 5
described above to generate the forecast for Hydro One Distribution load. 6
7
Filed: 2013-12-19
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 31 of 49
APPENDIX C 1
END-USE MODEL 2
3
The following briefly describes the methodology used in the end-use model. 4
5
Residential Sector 6
The residential energy forecast is determined by forecasting the number of accounts times 7
appliance saturation rates and unit energy consumption expressed in the following equation: 8
jijijijis UECSNUSE ,,,Re ** 9
Where 10
sUSERe is residential energy consumption 11
N is the number of residential accounts 12
S is the residential appliance saturation rate 13
UEC is the unit energy consumption per end use 14
I is the index for appliances (space heating, space cooling, water heater and base load) 15
J is the index for customer types—year-round residential customers and seasonal 16
residential customers 17
18
The following section describes each component of the equation in detail. 19
20
The base-year number of households is taken from Hydro One Distribution billing system. 21
The forecast in the growth of the number of residential accounts is based on a forecast of 22
housing starts. The number of residential accounts is the current number of residential 23
accounts plus the forecast of net additional accounts to be added each year. 24
The base-year end-use shares (space heating, water heating and air conditioning), and fuel 25
switching (space/water heating) information are based on the latest customer and 26
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 32 of 49
conservation survey results undertaken by Hydro One for year-round and seasonal 1
customers. 2
The trends for end-use shares and fuel switching over the forecasting period reflect the 3
provincial trends from the Hydro One provincial residential end-use model, as well as 4
information specific to Hydro One Distribution. 5
The base-year end-use UEC’s are based on the provincial residential end-use model with 6
adjustments for heating degree days, cooling degree days, income, household size, square 7
footage and household vintage. 8
9
Commercial Sector 10
The commercial energy forecast is based on the following equation: 11
USEcom =USEcom (-1) * (1+Expected annual growth rate ) 12
13
Where 14
USEcom is the commercial energy consumption for the forecast year 15
USEcom (-1) is the commercial energy consumption for the previous year. The base 16
year (2012) consumption is taken from the latest Hydro One Distribution billing system 17
corrected for abnormal weather effects 18
Expected annual growth rates are based on the Hydro One provincial commercial 19
end-use model. Where appropriate, the values are adjusted to reflect specific 20
distribution business characteristics. 21
The model uses an end-use framework to provide estimates of energy use by building 22
type. The building types include multi-residential, office, elementary and secondary 23
school, college and universities, health, public service, retail, grocery, accommodation, 24
recreation, religious/cultural, warehouse, commercial miscellaneous. non-building 25
related segments and streetlight. 26
27
Filed: 2013-12-19
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 33 of 49
Industrial Sector 1
The industrial energy forecast is based on the following equation: 2
USEind =USEind (-1) * (1+Expected annual growth rate) 3
Where 4
USEind is the industrial energy consumption for the forecast year 5
USEind (-1) is the industrial energy consumption for the previous year. The base year 6
(2012) consumption is taken from the latest Hydro One Distribution billing system 7
corrected for abnormal weather effects 8
Expected annual growth rates are based on the Hydro One provincial industrial end-9
use model. Where appropriate, the values are adjusted to reflect specific distribution 10
business characteristics. 11
The model uses an end-use framework to provide estimates of energy use by industrial 12
segments including Fishing, logging, Forestry Service, Mining, Petroleum, Food and 13
Beverage, Tobacco, Rubber and Plastic, Textile and Clothing, Wood and Furniture, 14
Paper and Printing, Primary Metal, Fabricated Metal Products, Transportation 15
Equipment, Electronics etc. 16
17
Agricultural Sector 18
The Agricultural sector forecast is based on the following equation: 19
USEagri =USEagri (-1) * (1+Expected annual growth rate) 20
Where 21
USEagri is the agricultural energy consumption for the forecast year 22
USEagri (-1) is the agricultural energy consumption for the previous year. The base 23
year (2012) consumption is taken from the latest Hydro One Distribution billing system 24
corrected for abnormal weather effects 25
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 34 of 49
Expected annual growth rates are based on the Hydro One provincial agricultural end-1
use model. Where appropriate, the values are adjusted to reflect specific distribution 2
business characteristics. 3
The model uses an end-use framework to provide estimates of energy use by 4
agricultural segments including Animal Production, Fruit and Vegetable Farming, 5
Grain Farming, Green Housing and Floriculture and Miscellaneous etc. 6
7
8
Filed: 2013-12-19
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 35 of 49
APPENDIX D 1
WEATHER NORMALIZATION FOR TOTAL LOAD AND BY RATE CLASS 2
3
Weather Normalization for Total Utility Load 4
5
Hydro One’s weather normalization methodology for total utility load is summarized as 6
follows: 7
8
An equation relating daily energy and daily weather conditions is developed using the 9
latest 4 years of data. This time frame allows the analysis to reflect the most recent load 10
mix while having sufficient data to quantify its weather sensitivity. For example, the 11
share of space cooling energy relative to total energy has increased rapidly over the past 12
decade; using too long a time series of historical data may lead to significant under-13
estimation of the weather sensitivity of load in the summer. 14
To better isolate the impact of weather, systematic changes in daily loads are identified 15
and filtered out before the regression analysis begins. The systematic effects removed 16
include growth trends, cyclical variations, day-of-the-week effects and holiday effects. 17
The objective is to filter the data to weather-related load and noise (random effect). 18
Different types of weather data are used in the analysis. For winter loads, weather data 19
include temperature, wind speed and cloud opacity. For summer loads, weather data 20
include temperature, humidity and cloud opacity. Because weather effects cumulate over 21
several days, the temperatures for the current day as well as the previous 3 or 4 days are 22
also used as explanatory variables in the model. The relationship between energy and 23
weather may be represented by the following function: 24
25
Weather- Related Energy = f (Weather Conditions) + Random Term (1) 26
27
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 36 of 49
where the random term reflects any remaining variations that are not explained 1
systematically by weather. The random term is assumed to be distributed independently, 2
identically and normally with mean equals to zero. 3
4
The coefficients from Equation (1) are estimated using the most recent 4 years of daily 5
load and weather data. These coefficients indicate the sensitivity of load in the service 6
territory relative to today’s temperature, yesterday’s temperature and all other weather 7
variables included in the equation. The estimated coefficients are multiplied by the actual 8
weather data for the corresponding weather variable in the equation to determine the 9
estimated weather-related energy for the day. This process is repeated for each day of the 10
period for which weather-correction is performed. 11
12
Estimated Weather-Related Energy = f (Actual Weather Conditions and Estimated 13
Coefficients) (2) 14
15
Equation (2) is used to determine what “normal” weather-related loads would be for each 16
day of the year given the current mix of weather-sensitive loads in that service territory. 17
This is done by running the equation with each of the last 31 years of daily weather data 18
for that day plus the seven days on either side of it. The average of the estimated 19
weather-related loads for the 15 days times 31 years (465 observations) is deemed to be 20
the “normal” weather-related energy for that day. Using 31 years of weather history is 21
considered adequate to approximate normal weather. 22
23
Normal Weather-Related Energy (for each day) = Average (31 years of Estimated 24
Weather-Related Energy for that Day +/- 7 Days) 3) 25
26
Filed: 2013-12-19
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 37 of 49
On a daily basis, the weather correction is derived as the difference between the estimated 1
and normal weather- related energy: 2
3
Weather Correction for Energy = Normal Weather-Related Energy – Estimated Weather-4
Related Energy (4) 5
6
Weather-corrected energy is defined to be actual energy plus the weather correction in 7
any given period. For any period that is more than one day (e.g., a month), the total 8
weather correction is the sum of the daily weather correction. 9
10
Weather-Corrected Energy = Actual Energy + Weather Correction for Energy (5) 11
12
For example, a summer day for which the combination of temperature and humidity are 13
above normal yields a negative weather correction. The weather correction in this case 14
should be viewed as the amount to be subtracted from the above normal actual to get the 15
weather-corrected energy. Similarly, a warm winter day would have a positive weather 16
correction as the weather corrected value for that day should be higher than the below 17
normal actual. 18
19
Weather Normalization by Rate Class 20
21
Weather correction by rate class is derived from weather correction for the total utility using 22
the electric space heating and cooling shares by rate class or segment as detailed below. 23
24
Weather correction for the total load is discussed above using daily energy for the utility. 25
The amount of weather correction is measured on a daily basis. 26
27
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 38 of 49
Using average daily temperature for each day, the daily weather correction is grouped 1
into “weather correction for space heating” and “weather correction for space cooling”. 2
For example, if average daily temperature is -1, the weather correction for that day is 3
allocated to “weather correction for space heating” load. The daily weather correction 4
results are aggregated into annual or monthly weather correction estimates. 5
6
Using load shape analysis and residential appliance saturation estimates, the amount of 7
space heating and cooling load over a year or month are estimated for each rate class. 8
Next, for each rate class, the cooling and heating weather correction amount are 9
calculated using the total cooling and heating weather correction amount multiplied by 10
the corresponding cooling and heating shares. The weather-corrected load for each rate 11
class is estimated by adding the weather correction estimates by rate class to the 12
corresponding (actual) load for each rate class. 13
14
Filed: 2013-12-19
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 39 of 49
APPENDIX E 1
STATISTICAL APPENDIX 2
3
4
Table E.1 5
6
Comparison of Forecasts for Previous Rate Submissions with Actual 7
(GWh) 8
9
10
11
12
2005 2007 2009 Weather % Difference from Weather Corrected Actual
Forecast Forecast Forecast Corrected 2005 2007 2009
Year (EB-2005-0378) (EB-2007-0681) (EB-2009-0096) Actual Actual Forecast Forecast Forecast
2005 23,027 22,969 23,182 0.25
2006 22,950 22,921 22,485 0.13
2007 23,074 22,945 22,966 22,909 0.47 -0.09
2008 23,062 22,845 22,624 0.95
2009 23,029 22,629 22,660 22,299 1.62 -0.14
2010 22,007 22,062 21,977 -0.25
2011 21,851 22,023 21,718 -0.78
3-Year Average 0.28 0.83 -0.39
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 40 of 49
Table E.2 1
Consensus Forecast for Ontario GDP and Housing Starts2
3
Survey of Ontario GDP Forecast (annual growth rate in %)
2014 2015 2016 2017 2018 2019Global Insight (Feb 2014) 2.3 2.7 2.7 2.6 2.4 2.8Conference Board (Feb 2014) 2.1 2.4 2.5 2.4 2.2U of T (Jan 2014) 2.2 3.0 2.9 2.9 2.5 2.3C4SE (Jan 2014) 2.2 2.7 2.5 2.9 2.8 2.1CIBC (Oct 2013) 2.3BMO (Feb 2014) 2.3 2.5RBC (Dec 2013) 2.6 2.9Scotia (Feb 2014) 2.0 2.2TD (Jan 2014) 2.2 2.6Desjardins (Feb 2014) 2.2 2.7Central 1 (Oct 2013) 1.9 2.6 2.8 3.2 3.2National Bank (Dec 2013) 2.1 2.5Laurentian Bank (Jan 2014) 2.3 2.5 Average 2.2 2.6 2.7 2.8 2.6 2.4
Survey of Ontario Housing Starts Forecast (in 000's)
2014 2015 2016 2017 2018 2019Global Insight (Feb 2014) 58.5 59.4 59.4 59.4 59.1 57.8Conference Board (Feb 2014) 61.3 62.2 69.7 77.6 85.5U of T (Jan 2014) 61.7 63.7 66.7 67.6 68.4 69.3C4SE (Jan 2014) 68.4 75.2 73.7 75.2 79.0 80.4CIBC (Oct 2013) 56.0BMO (Feb 2014) 55.0 55.0RBC (Dec 2013) 60.3 58.0Scotia (Feb 2014) 56.0 54.0TD (Jan 2014) 57.0 52.0Desjardins (Feb 2014) 58.2 58.8Central 1 (Oct 2013) 61.1 68.3 74.5 80.5 84.7National Bank (Dec 2013) 58.0 64.5Laurentian Bank (June 2013) 55.0 53.0 Average 59.0 60.3 68.8 72.1 75.3 69.2
Forecast updated on March 4, 2014
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 41 of 49
Table E.3 1
Economic Variables for Ontario 2
3
4
5 6
7
YearGDP
(2002 M$)
%
Change
Population
(1,000's)
%
Change
Housing
(1,000's)%Change
2003 484,341 1.4 12,242 1.3 84.7 1.3
2004 496,208 2.5 12,391 1.2 85.1 0.5
2005 510,626 2.9 12,528 1.1 78.2 -8.2
2006 522,845 2.4 12,665 1.1 74.2 -5.1
2007 529,913 1.4 12,791 1.0 67.6 -9.0
2008 529,132 -0.1 12,932 1.1 75.1 11.2
2009 512,812 -3.1 12,998 0.5 49.7 -33.8
2010 530,242 3.4 13,135 1.1 60.7 21.9
2011 541,728 2.2 13,264 1.0 67.7 11.6
2012 548,997 1.3 13,412 1.1 77.1 14.0
2013 555,854 1.2 13,538 0.9 61.0 -20.9
2014 568,126 2.2 13,667 1.0 59.0 -3.3
2015 582,925 2.6 13,806 1.0 60.3 2.2
2016 598,547 2.7 13,944 1.0 68.8 14.1
2017 615,307 2.8 14,089 1.0 72.1 4.8
2018 631,428 2.6 14,240 1.1 75.3 4.4
2019 646,582 2.4 14,390 1.1 69.2 -8.1
Filed: 2013-12-19
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 42 of 49
Table E.4 1
Mid-Year Number of Customers Before Harmonization 2
3
Rate Class 2008 2009
Urban Residential 86,149 87,416
High Density Residential 316,550 322,314
Low Density Residential 278,202 280,137
High Density Seasonal 71,347 71,237
Low Density Seasonal 84,622 84,889
Farm 1-Phase 87,442 87,063
Farm 3-Phase 1,267 1,271
Urban General service 6,165 7,236
General Service 1-Phase 62,400 61,724
General Service 3-Phase 16,758 16,382
Transmission 371 371
Lighting 37,561 37,864
Aquired Residential 146,424 147,089
Acquired General Service 22,103 22,161
Acquired Large Users 7 7
Total 1,217,369 1,227,161
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 43 of 49
1
Table E.4 Continued: Mid-Year Number of Customers After Harmonization 2
3
Rate Class 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
Dgen 88 106 248 477 731 1,010 1,289 1,523 1,719 1,908
GSd 7,028 7,183 6,550 6,669 6,748 6,113 6,196 6,282 6,375 6,458
GSe 98,010 98,095 98,513 98,568 98,785 93,508 93,788 94,081 94,493 94,764
R1 393,658 402,173 403,304 409,901 415,301 438,279 443,872 449,678 456,145 461,880
R2 366,295 368,479 370,995 373,980 376,395 335,043 337,693 340,418 343,578 346,199
Seasonal 158,247 157,017 153,653 153,253 153,677 143,666 144,188 144,731 145,457 145,966
ST 798 794 795 800 805 810 816 822 829 835
UGd 1,266 1,272 1,185 1,184 1,189 1,901 1,907 1,913 1,921 1,927
UGe 11,641 11,650 12,308 12,307 12,343 17,768 17,808 17,851 17,919 17,962
UR 156,008 159,086 167,672 169,795 171,883 209,540 211,691 213,918 216,443 218,631
STR 4737 4771 4,724 4,804 4,845 4,883 4,927 4,973 5,026 5,071
SEN 33692 31447 30,504 30,380 30,227 30,009 29,840 29,671 29,554 29,391
USL 5,498 5,504 5,512 5,562 5,612 5,642 5,691 5,734 5,776 5,824
Total 1,236,965 1,247,577 1,255,963 1,267,680 1,278,541 1,288,172 1,299,705 1,311,594 1,325,235 1,336,816
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 44 of 49
Table E.5 1
2
Hydro One Distribution Load History and Forecast in GWh 3
4
5
Year Actual/Forecast MWh Growth Normalized Weather MWh Growth
2008 39,727 -1.8 40,036 -1.2
2009 38,672 -2.7 39,282 -1.9
2010 37,980 -1.8 38,247 -2.6
2011 37,641 -0.9 38,062 -0.5
2012 37,627 0.0 37,419 -1.7
2013 38,331 1.9 37,496 0.2
2014 37,508 -2.1 37,508 0.0
2015 37,620 0.3 37,620 0.3
2016 37,824 0.5 37,824 0.5
2017 38,108 0.7 38,108 0.7
2018 38,111 0.0 38,111 0.0
2019 37,961 -0.4 37,961 -0.4
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 45 of 49
Table E.6 1
2
Actual Sales by Rate Class 3
Before Rate Harmonization in GWh 4
5
6
7
8
Table E.6 Continued: Actual Sales 9
and Forecast After Rate Harmonization in GWh 10
11
12
13
Rate Class 2008 2009Urban Residential 820 810High Density Residential 3,736 3,711Normal Density Residential 3,851 3,789High Density Seasonal 340 332Low Density Seasonal 398 387Farm - 1 Phase 1,833 1,774Farm - 3 Phase 222 218Urban General service 569 569General Service - 1Phase 1,133 1,104General Service - 3 Phase 2,983 2,974Transmission 1,089 1,090Lighting 148 146Aquired Residential 1,469 1,437Acquired General Service 2,071 2,032Acquired Large Users 299 300Embedded 16,541 15,834Total 37,502 36,509
Rate Class 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
Dgen 6 8 11 15 19 21 23 24 25 26
GSd 2,925 3,100 2,888 2,777 2,777 2,404 2,438 2,469 2,472 2,464
GSe 2,269 2,306 2,518 2,401 2,382 2,195 2,206 2,216 2,199 2,172
R1 4,393 4,402 4,396 4,544 4,574 5,052 5,109 5,184 5,228 5,255
R2 5,494 5,491 5,515 5,629 5,592 4,933 4,923 4,933 4,912 4,872
Seasonal 718 701 666 676 668 474 471 473 473 469
ST 17,000 16,787 17,082 16,497 16,532 16,560 16,629 16,731 16,718 16,637
UGd 677 686 677 650 648 1,068 1,077 1,086 1,082 1,073
UGe 400 397 415 398 396 604 609 613 611 605
UR 1,541 1,541 1,563 1,612 1,621 2,001 2,016 2,039 2,050 2,053
STL 126 125 127 122 123 124 125 125 126 127
SEN 20 19 19 22 22 22 22 22 22 22
USL 23 23 23 23 23 24 25 25 26 26
Total 35,593 35,587 35,901 35,365 35,378 35,483 35,674 35,940 35,943 35,801
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 46 of 49
Table E.7 1
2
Weather Corrected Sales 3
Before Rate Harmonization in GWh 4
5
6
7
8
Table E.7 Continued: Weather Corrected Sales 9
and Forecast After Rate Harmonization in GWh 10
11
12
13
14
Rate Class 2008 2009Urban Residential 829 825High Density Residential 3,778 3,778Normal Density Residential 3,895 3,858High Density Seasonal 344 338Low Density Seasonal 403 394Farm - 1 Phase 1,850 1,802Farm - 3 Phase 224 222Urban General service 575 578General Service - 1Phase 1,144 1,122General Service - 3 Phase 3,011 3,021Transmission 1,089 1,090Lighting 148 146Aquired Residential 1,499 1,488Acquired General Service 2,091 2,069Acquired Large Users 302 305Embedded 16,626 16,075Total 37,808 37,112
Rate Class 2010 2011 2012 2013 2014 2015 2016 2017 2018 2019
Dgen 6 8 11 15 19 21 23 24 25 26
GSd 2,926 3,150 2,959 2,777 2,777 2,404 2,438 2,469 2,472 2,464
GSe 2,270 2,343 2,580 2,401 2,382 2,195 2,206 2,216 2,199 2,172
R1 4,421 4,466 4,495 4,544 4,574 5,052 5,109 5,184 5,228 5,255
R2 5,529 5,571 5,640 5,629 5,592 4,933 4,923 4,933 4,912 4,872
Seasonal 722 711 681 676 668 474 471 473 473 469
ST 17,183 16,901 16,427 16,497 16,532 16,560 16,629 16,731 16,718 16,637
UGd 677 697 694 650 648 1,068 1,077 1,086 1,082 1,073
UGe 400 404 425 398 396 604 609 613 611 605
UR 1,551 1,563 1,599 1,612 1,621 2,001 2,016 2,039 2,050 2,053
STL 126 125 127 122 123 124 125 125 126 127
SEN 20 19 19 22 22 22 22 22 22 22
USL 23 23 23 23 23 24 25 25 26 26
Total 35,855 35,982 35,680 35,365 35,378 35,483 35,674 35,940 35,943 35,801
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 47 of 49
Table E.8a 1
2
Actual and 3
Forecast for Billing Peak in kW 4
5
6
7
8
9
10
11
12
Rate Class DGEN GSd Ugd ST Total
2008 66,624 10,549,230 1,830,892 35,182,285 47,629,031
2009 67,788 10,542,400 1,943,057 35,980,901 48,534,146
2010 59,361 10,288,535 1,981,526 36,362,897 48,692,319
2011 68,282 10,331,311 1,964,583 35,730,299 48,094,476
2012 81,512 10,050,244 1,912,569 36,409,471 48,453,796
2013 152,340 9,459,968 1,796,219 35,229,815 46,638,342
2014 184,998 9,459,635 1,792,366 35,656,983 47,093,982
2015 208,605 8,190,437 2,952,212 35,979,010 47,330,264
2016 227,242 8,306,523 2,978,661 35,937,113 47,449,539
2017 236,553 8,411,470 3,001,926 36,051,950 47,701,900
2018 246,017 8,421,313 2,991,451 35,823,052 47,481,834
2019 254,877 8,393,910 2,967,066 35,539,737 47,155,590
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 48 of 49
Table E.8b 1
2
Weather Corrected Actual 3
and Forecast for Billing Peak in kW 4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
Rate Class DGEN GSd Ugd ST Total
2008 66,342 10,504,548 1,823,137 34,744,764 47,138,791
2009 69,646 10,831,349 1,996,313 36,882,262 49,779,570
2010 56,860 9,854,946 1,898,019 34,830,459 46,640,284
2011 66,297 10,030,850 1,907,448 34,691,170 46,695,764
2012 80,371 9,909,510 1,885,788 35,862,030 47,737,698
2013 152,340 9,459,968 1,796,219 35,229,815 46,638,342
2014 184,998 9,459,635 1,792,366 35,656,983 47,093,982
2015 208,605 8,190,437 2,952,212 35,979,010 47,330,264
2016 227,242 8,306,523 2,978,661 35,937,113 47,449,539
2017 236,553 8,411,470 3,001,926 36,051,950 47,701,900
2018 246,017 8,421,313 2,991,451 35,823,052 47,481,834
2019 254,877 8,393,910 2,967,066 35,539,737 47,155,590
Updated: 2014-05-30
EB-2013-0416
Exhibit A
Tab 16
Schedule 2
Page 49 of 49
Table E.9 1
Hydro One Distribution CDM Impacts (GWh) by Rate Class 2
3
Note: All savings are at end-use level 4
Rate Class 2013 2014 2015 2016 2017 2018 2019
R1 212 227 263 277 282 348 430
R2 265 283 265 279 284 350 433
UR 77 82 107 113 115 142 175
Seasonal 33 35 25 27 27 34 42
GSe 233 236 217 218 214 231 257
UGe 38 39 59 59 58 62 69
GSd 265 268 229 230 226 244 271
UGd 62 63 103 104 101 110 122
ST 154 156 157 158 156 169 189
Total 1,339 1,389 1,425 1,466 1,463 1,688 1,988